Please give me some hints. What I'm trying to accomplish is to port CM10 and 10.1 and kernel 3.4 to LG E400. E400 has CM9 but the kernel is old - 2.6. I'm using kernel 3.4 for LG L3 II which has the same hardware.
I've compiled CM 10 but when I boot it with 3.4 kernel, I have a blank screen. I looked at dmesg and logcat, but I didn't figure what went wrong. I tried to boot CM 10 with 2.6 kernel but I have the same issue. I even try to test the 3.4 kernel with CM9 and it's the same story.
Can you give me some tips, how to figure out what went wrong?
dmesg for CM 10 using kernel 3.4: http://pastebin.com/M4R5tixn
logcat for CM 10 using kernel 3.4: http://pastebin.com/qRKttGTY
dmesg for CM 9 using kernel 3.4: http://pastebin.com/nppGkKB0
Thanks.
Here are some logcats for CM 10, kernel 3.4:
a. with blobs extracted from phone running CM9: http://pastebin.com/NKeYgEBw
b. with some blobs from LG L5 and adreno drivers for JB 4.1 from qualcomm site: http://pastebin.com/6D4VQDRa
c. with software renderer: http://pastebin.com/rtfiZi5s
d. with some blobs from LG L5 and latest adreno drivers from qualcomm site: http://pastebin.com/bcC6KCwj
In b. it seems adreno libs are loaded, surfaceflinger is running, so why I can't see android boot logo? All I'm able to see is a white background.
Using software renderer I see nothing but a blank display too.
dmesg says: init: could not import file /system/etc/init.local.rc, might this be relevant? Init.rc and init.board.rc are from CM 9.
edit1:
Since it didn't boot after I flashed the rom, I used ramdisk from CM9 and kernel 3.4 packing that into a boot.img and using that to boot.
Maybe using the ramdisk from CM9 with CM10 is why I am having problems?
edit2:
It seems zygote is crashing and restarting. Probably because of faulty init.rc.
edit3:
I fixed zygote crashing, now I have this:
Code:
E/memalloc( 128): /dev/pmem: Failed to open pmem device: No such file or directory
E/memalloc( 128): /dev/pmem: failed to initialize pmem area
E/gralloc ( 128): gralloc failed err=No such file or directory
It's weird there's no /dev/pmem because I've compiled the kernel with CONFIG_ANDROID_PMEM=y.
edit 4:
in board-xxx.c there's this comment:
Code:
/*LGE_CHANGE_S[[email protected]]20121102:
*Migrating QCT patch to remove pmem and fmem support*/
Also, there's no other reference to pmem in the file. It seems LG removed support for pmem. How do I get around this? gralloc needs /dev/pmem and there's no /dev/pmem because LG patched the kernel...
Code:
E/memalloc( 130): /dev/pmem: Failed to open pmem device: No such file or directory
E/memalloc( 130): /dev/pmem: failed to initialize pmem area
E/gralloc ( 130): gralloc failed err=No such file or directory
W/GraphicBufferAllocator( 130): alloc(240, 320, 2, 00000b03, 0 ...) failed -2 (No such file or directory)
E/SurfaceFlinger( 130): GraphicBufferAlloc::createGraphicBuffer(w=240, h=320) failed (No such file or directory), handle=0x0
E/BufferQueue( 130): [<pid_249>] dequeueBuffer: SurfaceComposer::createGraphicBuffer failed
Should I try to patch board-xxx.c to use pmem or should I try to compile CM10 with ION support? If so, how do I enable ION support before compiling CM10? adreno drivers support ION?
It's enough to add TARGET_USES_ION := true to BoardConfigCommon.mk ?
I think hardware/qcom/display/* might use only pmem in CM10. For CM10.1 I had to replace hardware/qcom/display with the one in CM10 since I had tons of compile errors...
I really wish i knw watever u r saying above... i would lke to help. By d way thnx for such great work. I knw u ll solve out eventually.
Sent from my LG-E400 using xda app-developers app
Your kernel is the problem.
Hi Guys,
I'm trying to port to Android 5.0 to a new device (BQ Aquaris E5 4G) and I can't make it fully boot. I'm always hitting this error :
Code:
Code:
pid: 634, tid: 634, name: servicemanager >>> /system/bin/servicemanager <<<
signal 6 (SIGABRT), code -6 (SI_TKILL), fault addr --------
I can't give you ADB Logcat as logcat doesn't seams to work (adb works fine, I can browse adb shell and so on). It kept giving me this error : "logcat read failure: No such file or directory"
The only thing I've got is some tombstone files witch all look similar (please see one attached)
If you could give me an hint that would be great
I tried to change the kernel also but everytime I boot it (using fastboot boot newboot.img) I get a dark blinking screen.
Thanks and merry christmas!
Hi guy i'm building LineageOS for Razer Forge TV. The rom works fine but i can't get DRM/Widevine to work.
The board for this device it's apq8084 like shamu, quark and kccat6.
I succesfuly build the vendor tree including this blobs:
Code:
/vendor/lib/libdrmfs.so
/vendor/lib/libdrmtime.so
/vendor/lib/liboemcrypto.so
/vendor/lib/libprdrmdecrypt.so
/vendor/lib/libQSEEComAPI.so
/vendor/lib/libwvdrm_L1.so
/vendor/lib/libWVStreamControlAPI_L1.so
/vendor/lib/drm/libdrmprplugin.so
/vendor/lib/drm/libdrmwvmplugin.so
/vendor/lib/mediadrm/libprmediadrmdecrypt.so
/vendor/lib/mediadrm/libprmediadrmplugin.so
/vendor/lib/mediadrm/libwvdrmengine.so
But nothing seems to work. Boardconfig.mk contains:
Code:
# DRM Protected Video
BOARD_USES_LIBDRM := true
BOARD_WIDEVINE_OEMCRYPTO_LEVEL := 1
I also try using blobs from shamu as many other roms similars for this board, with no luck.
As far i see in logcat the device can't get or create certs and credentials, so it's keep falling back to L3.
With L3 i can't get TV Netflix or casting to work...
Here is the logcat: https://pastebin.com/F43RkkKa
Any help or clue will be appreciated!
Thanks
Well this is hard to solve...as far i see in the logs in some point QSEECOMAPI or OEMCrypto fails to encrypt (or decrypt?) probably certs...
I test it all, even with bullhead blobs, no luck. I also replace the proprietary kestore.qcom.so with the opensource provided by LineageOS and i get the same error....
Code:
WVCdm : Could not read /data/mediadrm/IDM1013/ay64.dat2: No such file or directory
QSEECOMAPI: : Error::send command ioctl failed. ret = -1, errno = 22
DrmWidevineDash: Error: OEMCrypto_Initialize ioctl returns -1
WVCdm : DeviceFiles::RetrieveHashedFile: /data/mediadrm/IDM1013/L3/cert.bin does not exist
Good evening people.Today i got my Amazfit delivered to my home and i am REALLY excited!
i finally have my modular clock!
My idea is to create a github organization and gather all dev in one place.
Do you like to dev ?
Are you interested to make roms and other root stuff ?
Do you like designing watch faces ?
Lets gather em all!...If you are interested please tell me your email and i will add u as admin to the organization.
The organization that i just created:
https://github.com/amazfitDev
thanks for all!
Being an Recognized Developer i know how android stuff works. I am really interested on that clock and i would like to dev on that. Any other dev around ?
From the image on your GitHub, I assume you have a Pace. There's been a few development projects surrounding it already:
Modified version of Amazfit app with many features: https://github.com/edotassi/AmazMod
Custom Springboard pages: https://github.com/KieronQuinn/AmazfitSpringboardPluginExample
→ Music page: https://github.com/KieronQuinn/AmazfitSpringboardMusic
→ Calendar page: https://github.com/GreatApo/AmazfitPaceCalendarWidget
→ Rearrange pages: https://github.com/KieronQuinn/AmazfitSpringboardSettings
Phone > Watch > Phone communication in apps: https://github.com/KieronQuinn/AmazfitCommunication
→ Internet Companion: https://github.com/KieronQuinn/AmazfitInternetCompanion
→ Train Times: https://forum.xda-developers.com/smartwatch/amazfit/app-uk-train-times-live-train-t3779955/page1
Custom watchfaces: https://forum.xda-developers.com/sm...t-huamiwatchfaces-based-custom-t3760814/page1
Step notify app: https://github.com/KieronQuinn/AmazfitStepNotify
My GitHub username is obvious from the links but the email linked to that account is quinny898 (at) gmail (dot) com
Quinny899 said:
From the image on your GitHub, I assume you have a Pace. There's been a few development projects surrounding it already:
Modified version of Amazfit app with many features: https://github.com/edotassi/AmazMod
Custom Springboard pages: https://github.com/KieronQuinn/AmazfitSpringboardPluginExample
→ Music page: https://github.com/KieronQuinn/AmazfitSpringboardMusic
→ Calendar page: https://github.com/GreatApo/AmazfitPaceCalendarWidget
→ Rearrange pages: https://github.com/KieronQuinn/AmazfitSpringboardSettings
Phone > Watch > Phone communication in apps: https://github.com/KieronQuinn/AmazfitCommunication
→ Internet Companion: https://github.com/KieronQuinn/AmazfitInternetCompanion
→ Train Times: https://forum.xda-developers.com/smartwatch/amazfit/app-uk-train-times-live-train-t3779955/page1
Custom watchfaces: https://forum.xda-developers.com/sm...t-huamiwatchfaces-based-custom-t3760814/page1
Step notify app: https://github.com/KieronQuinn/AmazfitStepNotify
My GitHub username is obvious from the links but the email linked to that account is quinny898 (at) gmail (dot) com
Click to expand...
Click to collapse
amazfitmod. is it rom ? or mod ? for flashing rule ?
brmbjn said:
amazfitmod. is it rom ? or mod ? for flashing rule ?
Click to expand...
Click to collapse
It's a modded CN Amazfit app, with support for Strava and other features.
As Quinny mentioned, there are some devs around, including the the ones providing some multilingual ROMs. Most of us have a github account (this is my project).
Και στο Facebook μπορείς να μας βρεις
There are many reputable developers here in one place...
I apologize, maybe it's not a good question at this place, but:
what is the ADB command line that shows the QR code for on active Amazfit watch display?
gidano said:
There are many reputable developers here in one place...
I apologize, maybe it's not a good question at this place, but:
what is the ADB command line that shows the QR code for on active Amazfit watch display?
Click to expand...
Click to collapse
adb shell am start -n com.huami.watch.ble/com.huami.watch.qrcode.QrActivity
You need to be running as root first with "adb root"
https://gitlab.com/Neuer_User/PACEfied_AmazFit/wikis/Faq
Thank you very much! :highfive:
wow...
its true that there are many projects around...but why everything is separated ? we better get unified / under a same git org ?
i love all those actions...i am still playing with watchfaces though...still non-rooted.i want to investigate the hardware of the clock (MIPS device-chances for ROM build? etc)
wishing the best!
side said:
wow...
its true that there are many projects around...but why everything is separated ? we better get unified / under a same git org ?
i love all those actions...i am still playing with watchfaces though...still non-rooted.i want to investigate the hardware of the clock (MIPS device-chances for ROM build? etc)
wishing the best!
Click to expand...
Click to collapse
They're separated because this forum isn't segmented like device forums are, and nobody had set up an organisation until now.
I've been looking at the possibility of getting custom ROMs working, Huami won't release the kernel source so that limits us a lot, but using a generic MIPS build of 8.0 built for a IoT board (optimistic I know) I got it as far as graphics crashing because of an incompatible driver. There is a supposedly compatible open source driver for the GPU in the MESA project, but I don't have the knowledge to build it (or the space). Maybe you could help there?
Quinny899 said:
They're separated because this forum isn't segmented like device forums are, and nobody had set up an organisation until now.
I've been looking at the possibility of getting custom ROMs working, Huami won't release the kernel source so that limits us a lot, but using a generic MIPS build of 8.0 built for a IoT board (optimistic I know) I got it as far as graphics crashing because of an incompatible driver. There is a supposedly compatible open source driver for the GPU in the MESA project, but I don't have the knowledge to build it (or the space). Maybe you could help there?
Click to expand...
Click to collapse
the gallium3d stuff ? i think i can help there
side said:
the gallium3d stuff ? i think i can help there
Click to expand...
Click to collapse
Yes, in theory all I need is the driver itself and it may continue booting. I'd also like to get a device tree set up using the generic source as base, but no kernel source is annoying
Damn... you 2 speak your own language....
Right now I'm at home and not on a phone, here's a dump of everything I know and have got working on the Pace in terms of custom ROMs. I have a second device I use for testing, and it's currently got something I hacked together from this on it. I used that ROM's system image as base, and copied across the following files from Pacefied (that I can remember, it was a few weeks ago now):
/system/bin/logcatd
/system/lib/sdcard
I also used a compare program to copy across drivers and device specific files, and went about reading the logcat and copying missing things, getting it further and further into the boot. The graphics drivers (libEGL_VIVANTE) aren't compiled for Oreo, so they have namespace issues. After copying way too many files to /system/vendor/lib/egl (this alone is a massive red flag, they should not be there at all), it complains something being missing:
Code:
05-05 17:43:17.572 7503 7503 E vndksupport: Could not load /vendor/lib/egl/libEGL_VIVANTE.so from sphal namespace: dlopen failed: /system/vendor/lib/egl/libcutils.so: cannot locate "_memset16"....
05-05 17:43:17.572 7503 7503 E libEGL : load_driver(/vendor/lib/egl/libEGL_VIVANTE.so): unknown
05-05 17:43:17.573 7503 7503 F libEGL : couldn't find an OpenGL ES implementation
05-05 17:43:17.573 7503 7503 F libc : Fatal signal 6 (SIGABRT), code -6 in tid 7503 (surfaceflinger), pid 7503 (surfaceflinger)
05-05 17:43:17.671 7560 7560 I crash_dump32: obtaining output fd from tombstoned, type: kDebuggerdTombstone
05-05 17:43:17.671 7560 7560 E libc : failed to connect to tombstoned: No such file or directory
05-05 17:43:17.671 7560 7560 I crash_dump32: performing dump of process 7503 (target tid = 7503)
05-05 17:43:17.672 7560 7560 F DEBUG : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
05-05 17:43:17.672 7560 7560 F DEBUG : Build fingerprint: 'Android/aosp_ci20/ci20:8.1.0/OPM1.171019.011/mousew03081157:userdebug/test-keys'
05-05 17:43:17.672 7560 7560 F DEBUG : Revision: '00000001'
05-05 17:43:17.672 7560 7560 F DEBUG : ABI: 'mips'
05-05 17:43:17.672 7560 7560 F DEBUG : pid: 7503, tid: 7503, name: surfaceflinger >>> /system/bin/surfaceflinger <<<
05-05 17:43:17.672 7560 7560 F DEBUG : signal 6 (SIGABRT), code -6 (SI_TKILL), fault addr --------
05-05 17:43:17.679 7560 7560 F DEBUG : Abort message: 'couldn't find an OpenGL ES implementation'
05-05 17:43:17.679 7560 7560 F DEBUG : zr 00000000 at 00000001 v0 00000000 v1 bd35f1fe
05-05 17:43:17.679 7560 7560 F DEBUG : a0 00001d4f a1 00001d4f a2 00000006 a3 00000000
05-05 17:43:17.679 7560 7560 F DEBUG : t0 00000000 t1 00000001 t2 fffffffc t3 00000000
05-05 17:43:17.679 7560 7560 F DEBUG : t4 fffffffc t5 77eeb000 t6 00000000 t7 65704f20
05-05 17:43:17.680 7560 7560 F DEBUG : s0 7719c7e0 s1 00001d4f s2 00001d4f s3 7fa964c0
05-05 17:43:17.680 7560 7560 F DEBUG : s4 00000004 s5 00000001 s6 77825304 s7 77825048
05-05 17:43:17.680 7560 7560 F DEBUG : t8 00000000 t9 7709a380 k0 773a1e0b k1 00000000
05-05 17:43:17.680 7560 7560 F DEBUG : gp 7719c7e0 sp 7fa96488 s8 773a0000 ra 7709a6d0
05-05 17:43:17.680 7560 7560 F DEBUG : hi 00000000 lo 00000000 bva 77c77fe0 epc 7709a3c0
05-05 17:43:17.686 7560 7560 F DEBUG :
05-05 17:43:17.686 7560 7560 F DEBUG : backtrace:
05-05 17:43:17.686 7560 7560 F DEBUG : #00 pc 000193c0 /system/lib/libc.so (syscall+64)
05-05 17:43:17.686 7560 7560 F DEBUG : #01 pc 000196d0 /system/lib/libc.so (abort+164)
05-05 17:43:17.686 7560 7560 F DEBUG : #02 pc 00021e0b /system/lib/libEGL.so
So this driver isn't compatible, clearly. At this point I went and researched the GPU of the watch. It has a Vivante GC 1000, which has been used in a couple of Android devices before but they're ARM and ancient so are not helpful. However, the etnaviv project has created some support for it, I'm not sure how helpful it is though: https://github.com/etnaviv/etna_viv, it's also been merged into the mesa driver. Android x86 uses Mesa, so it's definitely compatible with Android, but we need a tree set up first as far as I know
That's where my experience ends. I've never built Android from source before, nor do I have the space for it at the moment. I'd be willing to assist in setting up a tree, though. The earlier linked CI20 code is here: https://github.com/MIPS/CI20_android
Also remembered this: I think this is the commit that's causing the driver to be incompatible: https://android.googlesource.com/platform/bionic/+/b0cf928, as _memset16 is in a different place than it's expected to be. Using a newer lib in the egl folder does not fix this.
@Quinny899
perfect findings mate. Everything seems really nice! and your try to make stuff working is really respected!....although i got some bad news .
device tree and kernel sources are mandatory...even if we build the device tree,there is no chance to get the kernel source code ...and there is no chance to build AOSP from source without the kernel source.
my opinion is to push huami release the kernel source. Since its Linux Kernel (there is no doubt about it ) its licensed as GPL. They have to release it. i can get some FOSS maniacs to push them
edit1:
This device tree ( https://github.com/MIPS/CI20_android/tree/dev-mips-oreo-mr1 ) could do the job but still...no kernel.although the device tree is not that hard to be built.i can do that.
For a flashable zip built by AOSP sources we need:
device tree + proprietary libs ( those u copied ) and a kernel .the rest could be unified or something
edit2:
Even if i got a flashable rom i cant test it on my self.i dont have 2nd watch and i wont risk it
edit3:
about the commit u said...its easy to be reverted.as i see its a standalone commit.doesnt depend on any other repo...really easy to be reverted but i dont think that reverting it fixes the driver problem. it may be fixed from runtime side but not kernel/module side
edit4:
about the log u posted.is it a last_kmsg ? or dmsg ? or simple log ?
whatever it is i can confirm its bionic's issue since libc blows up
The log is a snippet of a logcat, the kernel doesn't actually crash so the last message wouldn't be of use as far as I know
If you want to send any test stuff to me, I have a second device I can flash on (using dd - nice and dangerous) to see if it works
In terms of the lack of kernel, could we use the generic MIPS kernel from AOSP and just not flash the boot image maybe?
This is the response I got when I asked for the kernel source:
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Quinny899 said:
The log is a snippet of a logcat, the kernel doesn't actually crash so the last message wouldn't be of use as far as I know
If you want to send any test stuff to me, I have a second device I can flash on (using dd - nice and dangerous) to see if it works
In terms of the lack of kernel, could we use the generic MIPS kernel from AOSP and just not flash the boot image maybe?
This is the response I got when I asked for the kernel source:
Click to expand...
Click to collapse
hmmm about the generic kernel .i dont know...those "generic" solutions usually doesnt work but we can try it .keep in mind that a source-built kernel product is different than the one the vendor is providing. kernel and rom are strictly tided....there is no chance to build THE SAME kernel as the vendor's one. a source-built one needs patches to be bootable.
about their response.really ? if its linux kernel(which i am sure it is) thats against the law and they will have serious problems O_O'
side said:
hmmm about the generic kernel .i dont know...those "generic" solutions usually doesnt work but we can try it .keep in mind that a source-built kernel product is different than the one the vendor is providing. kernel and rom are strictly tided....there is no chance to build THE SAME kernel as the vendor's one. a source-built one needs patches to be bootable.
about their response.really ? if its linux kernel(which i am sure it is) thats against the law and they will have serious problems O_O'
Click to expand...
Click to collapse
A lot of Chinese vendors ignore the GPL, Xiaomi themselves aren't great at releasing sources either
Hello,
App used : com.iiordanov.freebVNC from PlayStore, project on github
The application crashes randomly with the following error, after opening several windows in VNC quickly.
APK Version : 4.8.0
Server Version : tigervnc-server-1.8.0-5.el7.x86_64
W/Bitmap: Called getHeight() on a recycle()'d bitmap! This is undefined behavior!
A/Bitmap: Error, cannot access an invalid/free'd bitmap here!
A/libc: Fatal signal 6 (SIGABRT), code -6 in tid 7606 (.iiordanov.bVNC), pid 7606 (.iiordanov.bVNC)
Do you have an idea ?
Thanks.